Study Aims

The overall aims are to investigate the distributions and correlates of problem gambling in young people, and to describe the developmental trajectories from early childhood of adolescents with problem gambling

Specific objectives:

1. To describe the prevalence of gambling behaviours at 17 years, and their relation to parental gambling patterns.

2. To describe the associations of problem gambling with other risk taking behaviours, alcohol and drug use in adolescence and with debt.

3. To elucidate the interactions between gambling behaviour, developmental conditions such as ADHD and affective states such as depression

4. To investigate the relationship between problem gambling behaviour in adolescents and impulsivity and altered decision-making.

5. To describe developmental trajectories from early childhood of adolescents with problem gambling.

6. In those with a parent who gambles, to investigate the resilience factors which are associated with the young person not developing problem gambling behaviours.

Study design

The Avon Longitudinal Study of Parents and Children is a population-based cohort from the South West of England. Allbirths to women resident in the former Avon Health Authority area, with an expected date of delivery between 1st April 1991 and 31st December 1992 were eligible for enrolment in the ALSPAC study, resulting in a total cohort of 14,062 live births. Active contact is maintained with over 10,000 participants and their parents. At the age of 17 years, the young people and each of their parents will be sent a postal questionnaire (with an option to complete questionnaire online) and the young person will be invited to a research clinic. We will assess the children's and their parents' gambling behaviour and associated cognition in the 17-year questionnaire, and assess the young person's decision-making using a computerised test in the 17-year clinic.

a) gambling behaviour

We propose to use questions from theCanadian Problem Gambling Inventory (CPGI) The CPGI consists of 31 questions: including 'core' items to give a prevalence rate for problem gambling, and other items which are indicators of gambling involvement and correlates of problem gambling to develop a profile of different types of gamblers. Using the CPGI, respondents are classified into five groups: Non-Gamblers, Non-Problem Gamblers, Low Risk Gamblers, Moderate Gamblers and Problem Gamblers. Although ALSPAC previously used the South Oaks Gambling Screen (SOGS) with the parents, the CPGI is preferred because it gives a better spread of categories of gambler.

Problem gambling will be defined using theProblem Gambling Severity Index (PGSI), a 9-item scale derived from the 31 items larger screen. The PGSI is a well validated test which was used in the 2007 BGPS (Wardle, 2007), and in other international prevalence studies (e.g., inAustralia). A validation study (Wenzel, 2004), comparing the performance of the SOGS, the PGSI and the Victorian Gambling Screen (VGS) found that the PGSI outperformed the other two screens. The PGSI items each have four response options. For each item, 'sometimes' is given a score of one, 'most of the time' scores two, and 'almost always' scores three. A score of between zero and 27 is therefore possible, and a threshold of 8 is used to identify problem gamblers., and a score of 3-7 defines 'moderate risk' of PG. Questions from the CPGI, including the PGSI, will be included in the 17-year questionnaires, to be sent to all young people still in the cohort and to both their parents, between January 2009 and June 2010.

b) gambling cognition and decision-making

We propose to use theGambling Related Cognitions Scale(GRCS; Raylu & Oei, 2004) to assess cognition in gambling. The GRCS is a five-factor 23-item questionnaire, which includes questions which cover the wide range of gambling-related cognitive errors that have been reported in the gambling literature. The GRCS asks participants to use a 7-point Likert scale to indicate the extent to which they agreed with the value expressed in each statement. Scoring consisted of totalling the values such that the higher the score the higher the number of GRC displayed. The GRCSwill be included in the 17-year questionnaires, to be sent to all young people and to their parents.

As well as collecting behavioural data to identify types of gamblers (PGSI), and attitudinal data to assess gamblers' cognitive style (GRCS), we believe it is also important to directly test the young people to objectively measure their decision making under gambling conditions. We propose to use a computerised version of The Iowa Gambling Task (IGT), a measure of decision taking under ambiguity. The IGT is based on the Somatic Marker hypothesis (Bechara et al, 1994), and gives information about the development/impairment of orbitofrontal/ventromedial prefrontal cortex. Low performances in IGT have been linked to pathological gambling, alcohol dependence, Parkinson's disease and ADHD (in both adults and children). The IGT can be done manually or computerised: it consists on four decks of cards, two of them bring big wins and big losses, and the other two bring small wins but small losses and in the long-term are advantageous. Participants have different levels of awareness though the test, most participants know and can explain which decks are "advantageous" at card 80 (100 cards possible). Every participant will perform one "round" of the computerized adult version of the IGT (100 cards). Estimated time 10-15 minutes. The computerised IGT will be applied to all ALSPAC participants in the 17-year+ clinic (expected n=6000) between October 2008 and September 2010.

Other measures

The ALSPAC study has extensive epidemiological data regarding parental socioeconomic status, co-morbidity and gambling history measured with the SOGS scale (measured when children were 7 years). The children's behavioural and psychological profile has been measured since infancy, with some measures such as the Strengths and Difficulties Questionnaire (SDQ) repeated at several time points during childhood and adolescence. Psychometric assessments, including measures of cognition, attention and depression in children have been collected at various time points. (see appendix for table of measures already collected).

At 15-16 yrs, the participants have been assessed in the 15+ clinic, with direct measures of IQ (WASI), impulsivity (Stop Signal) and psychopathology (Development and Well Being Assessment- DAWBA). Over 6000 participants were tested in the 15+ clinic.GCSE results on the whole cohort at 16 yrs will be made available by DCSF.

At the 17 year+ clinic, which starts in autumn 2008, young people will complete computerised assessments of mood, funded by a grant from the Wellcome Trust (PI Lewis).We will assess depression and anxiety disorders using the revised clinical interview schedule (CIS-R), that has been widely used in the Psychiatric Morbidity Surveys in the UK, and assess psychological factors, including depressogenic cognitive styles, that are associated with depression and appear to reflect underlying vulnerability. Negotiations are advanced with government departments to fund computerised questions on debt and attitudes to financial risk taking in the 17+ clinic.